R. Sedghi; Y. Abbaspour Gilandeh
Abstract
Suitable soil structure is important for crop growth. One of the main characteristics of soil structure is the size of soil aggregates. There are several ways of showing the stability of soil aggregates, among which the determination of the median weight diameter of soil aggregates is the most common ...
Read More
Suitable soil structure is important for crop growth. One of the main characteristics of soil structure is the size of soil aggregates. There are several ways of showing the stability of soil aggregates, among which the determination of the median weight diameter of soil aggregates is the most common method. In this paper, a method based on adaptive neuro fuzzy inference system (ANFIS) was used to describe the soil fragmentation for seedbed preparation with combination of primary and secondary tillage implements including subsoiler, moldboard plow and disk harrow. Adaptive neuro fuzzy inference system (ANFIS) is a suitable approach to solving non-linear problems. ANFIS is a combination of fuzzy inference system (FIS) and an artificial neural network (ANN) method and it uses the ability of both models. In this study, the model inputs included “soil moisture content”, “tractor forward speed”and “working depth”. The performance of the model was evaluated using the statistical parameters of root mean square error (RMSE), percentage of relative error (ε), mean absolute error (MAE) and the coefficient of determination (R2). These parameters were determined as 0.135, 3.6%, 0.122 and 0.981, respectively. For the evaluation of the ANFIS model, the predicted data using this model were compared to the data of artificial neural network model. The simulation results by ANFIS model showed to be closer to the actual data compared with those made by the artificial neural network model.
Y. Abbaspour Gilandeh; R. Sedghi
Abstract
In this study, a knowledge-based fuzzy logic system was developed on experimental data and used to predict the draft force and energy requirement of tillage operation. In comparison with traditional methods, the fuzzy logic model acts more effectively in creating a relationship between multiple inputs ...
Read More
In this study, a knowledge-based fuzzy logic system was developed on experimental data and used to predict the draft force and energy requirement of tillage operation. In comparison with traditional methods, the fuzzy logic model acts more effectively in creating a relationship between multiple inputs to achieve an output signal in a nonlinear range. Field experiments were carried out in a sandy loam soil on coastal plain at the Edisto Research and Education Center of Clemson University near Blackville, South Carolina (Latitude 33˚ 21"N, Longitude 81˚ 18"W). In this paper, a fuzzy model based on Mamdani inference system has been used. This model was developed for predicting the changes of draft force and energy requirement for subsoiling operation. This fuzzy model contains 25 rules. In this investigation, the Mamdani Max-Min inference was used for deducing the mechanism (composition of fuzzy rules with input). The center of gravity defuzzification method was also used for conversion of the final output of the system into a classic number. The validity of the presented model was achieved by numerical error criterion, based on empirical data. The prediction results showed a close relationship between measured and predicted values such that the mean relative error of measured and predicted values were 3.1% and 2.94% for draft resistant force and energy required for subsoiling operation, respectively. The comparison between the fuzzy logic model and the regression models showed that the mean relative errors from the regression model are greater than that from the fuzzy logic model.
